Add try_update_params method#260
Conversation
sea-bass
left a comment
There was a problem hiding this comment.
Apologies for the delay -- I'm not really on these repos much these days.
Given this is a change in behavior, I don't feel like I should have the authority to make that decision. I'm going to leave this up to @pac48 as an original author of the package.
|
This was introduced with #205, but as discussed in the linked PR I think that it should have a different behavior. We could also deprecated the current one and add the proposed method with a different name if you prefer. |
|
@sea-bass I think the change in behavior makes sense in this case. I think we should merge this. |
|
There are a couple other options I think we should consider here. 1. Make no change, rely on user library to check if
|
Isn't this what the
I'm not sure if deprecating the old method is necessary, as it was never advertised in the documentation. But I agree with your arguments and the suboptimal naming for the new behavior. @saikishor do you have time to implement that, or shall I open a new PR? |
|
@christophfroehlich I can take care of it. I just pushed some changes |
christophfroehlich
left a comment
There was a problem hiding this comment.
Great. Let us also add both methods to the README..
|
Thank you for the updates. @christophfroehlich since we added the new method as opposed to changing the ABI, I assume we should get this PR into the next release? I am going to work on that in the morning, sorry it is taking so long. |
Only return true when the parameters are modified and return false for others cases
I've also modified it to use the unique_lock as it is much better than try_lock on the mutex itself.
Related to ros-controls/ros2_controllers#1198 (review)